freebsd-skq/share/man/man7/arch.7
mhorne dc653b0998 arch(7): small corrections for RISC-V
Document that RISC-V supports multiple page sizes: 4K, 2M, and 1G.

RISC-V's long double is always 128-bits wide, therefore quad precision.

Mention __riscv_float_abi_soft, which can be used to differentiate between
riscv64 and riscv64sf in userland code.

MFC after:	3 days
2020-06-23 17:17:13 +00:00

491 lines
16 KiB
Groff

.\" Copyright (c) 2016-2017 The FreeBSD Foundation. All rights reserved.
.\"
.\" This documentation was created by Ed Maste under sponsorship of
.\" The FreeBSD Foundation.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd June 23, 2020
.Dt ARCH 7
.Os
.Sh NAME
.Nm arch
.Nd Architecture-specific details
.Sh DESCRIPTION
Differences between CPU architectures and platforms supported by
.Fx .
.Ss Introduction
This document is a quick reference of key ABI details of
.Fx
architecture ports.
For full details consult the processor-specific ABI supplement
documentation.
.Pp
If not explicitly mentioned, sizes are in bytes.
The architecture details in this document apply to
.Fx 11.0
and later, unless otherwise noted.
.Pp
.Fx
uses a flat address space.
Variables of types
.Vt unsigned long ,
.Vt uintptr_t ,
and
.Vt size_t
and pointers all have the same representation.
.Pp
In order to maximize compatibility with future pointer integrity mechanisms,
manipulations of pointers as integers should be performed via
.Vt uintptr_t
or
.Vt intptr_t
and no other types.
In particular,
.Vt long
and
.Vt ptrdiff_t
should be avoided.
.Pp
On some architectures, e.g.,
.Dv powerpc
and AIM variants of
.Dv powerpc64 ,
the kernel uses a separate address space.
On other architectures, kernel and a user mode process share a
single address space.
The kernel is located at the highest addresses.
.Pp
On each architecture, the main user mode thread's stack starts near
the highest user address and grows down.
.Pp
.Fx
architecture support varies by release.
This table shows the first
.Fx
release to support each architecture, and, for discontinued
architectures, the final release.
.Pp
.Bl -column -offset indent "Architecture" "Initial Release" "Final Release"
.It Sy Architecture Ta Sy Initial Release Ta Sy Final Release
.It aarch64 Ta 11.0
.It alpha Ta 3.2 Ta 6.4
.It amd64 Ta 5.1
.It arm Ta 6.0 Ta 12.x
.It armeb Ta 8.0 Ta 11.x
.It armv6 Ta 10.0
.It armv7 Ta 12.0
.It ia64 Ta 5.0 Ta 10.4
.It i386 Ta 1.0
.It mips Ta 8.0
.It mipsel Ta 9.0
.It mipselhf Ta 12.0
.It mipshf Ta 12.0
.It mipsn32 Ta 9.0
.It mips64 Ta 9.0
.It mips64el Ta 9.0
.It mips64elhf Ta 12.0
.It mips64hf Ta 12.0
.It pc98 Ta 2.2 Ta 11.x
.It powerpc Ta 6.0
.It powerpcspe Ta 12.0
.It powerpc64 Ta 6.0
.It riscv64 Ta 12.0
.It riscv64sf Ta 12.0
.It sparc64 Ta 5.0 Ta 12.x
.El
.Ss Type sizes
All
.Fx
architectures use some variant of the ELF (see
.Xr elf 5 )
.Sy Application Binary Interface
(ABI) for the machine processor.
All supported ABIs can be divided into two groups:
.Bl -tag -width "Dv ILP32"
.It Dv ILP32
.Vt int ,
.Vt long ,
.Vt void *
types machine representations all have 4-byte size.
.It Dv LP64
.Vt int
type machine representation uses 4 bytes,
while
.Vt long
and
.Vt void *
are 8 bytes.
.El
.Pp
Some machines support more than one
.Fx
ABI.
Typically these are 64-bit machines, where the
.Dq native
.Dv LP64
execution environment is accompanied by the
.Dq legacy
.Dv ILP32
environment, which was the historical 32-bit predecessor for 64-bit evolution.
Examples are:
.Bl -column -offset indent "powerpc64" "ILP32 counterpart"
.It Sy LP64 Ta Sy ILP32 counterpart
.It Dv amd64 Ta Dv i386
.It Dv powerpc64 Ta Dv powerpc
.It Dv mips64* Ta Dv mips*
.It Dv aarch64 Ta Dv armv6/armv7
.El
.Pp
.Dv aarch64
will support execution of
.Dv armv6
or
.Dv armv7
binaries if the CPU implements
.Dv AArch32
execution state, however
.Dv armv5
binaries aren't supported.
.Pp
On all supported architectures:
.Bl -column -offset -indent "long long" "Size"
.It Sy Type Ta Sy Size
.It short Ta 2
.It int Ta 4
.It long Ta sizeof(void*)
.It long long Ta 8
.It float Ta 4
.It double Ta 8
.El
.Pp
Integers are represented in two's complement.
Alignment of integer and pointer types is natural, that is,
the address of the variable must be congruent to zero modulo the type size.
Most ILP32 ABIs, except
.Dv arm ,
require only 4-byte alignment for 64-bit integers.
.Pp
Machine-dependent type sizes:
.Bl -column -offset indent "Architecture" "void *" "long double" "time_t"
.It Sy Architecture Ta Sy void * Ta Sy long double Ta Sy time_t
.It aarch64 Ta 8 Ta 16 Ta 8
.It amd64 Ta 8 Ta 16 Ta 8
.It armv6 Ta 4 Ta 8 Ta 8
.It armv7 Ta 4 Ta 8 Ta 8
.It i386 Ta 4 Ta 12 Ta 4
.It mips Ta 4 Ta 8 Ta 8
.It mipsel Ta 4 Ta 8 Ta 8
.It mipselhf Ta 4 Ta 8 Ta 8
.It mipshf Ta 4 Ta 8 Ta 8
.It mipsn32 Ta 4 Ta 8 Ta 8
.It mips64 Ta 8 Ta 8 Ta 8
.It mips64el Ta 8 Ta 8 Ta 8
.It mips64elhf Ta 8 Ta 8 Ta 8
.It mips64hf Ta 8 Ta 8 Ta 8
.It powerpc Ta 4 Ta 8 Ta 8
.It powerpcspe Ta 4 Ta 8 Ta 8
.It powerpc64 Ta 8 Ta 8 Ta 8
.It riscv64 Ta 8 Ta 16 Ta 8
.It riscv64sf Ta 8 Ta 16 Ta 8
.El
.Pp
.Sy time_t
is 8 bytes on all supported architectures except i386.
.Ss Endianness and Char Signedness
.Bl -column -offset indent "Architecture" "Endianness" "char Signedness"
.It Sy Architecture Ta Sy Endianness Ta Sy char Signedness
.It aarch64 Ta little Ta unsigned
.It amd64 Ta little Ta signed
.It armv6 Ta little Ta unsigned
.It armv7 Ta little Ta unsigned
.It i386 Ta little Ta signed
.It mips Ta big Ta signed
.It mipsel Ta little Ta signed
.It mipselhf Ta little Ta signed
.It mipshf Ta big Ta signed
.It mipsn32 Ta big Ta signed
.It mips64 Ta big Ta signed
.It mips64el Ta little Ta signed
.It mips64elhf Ta little Ta signed
.It mips64hf Ta big Ta signed
.It powerpc Ta big Ta unsigned
.It powerpcspe Ta big Ta unsigned
.It powerpc64 Ta big Ta unsigned
.It riscv64 Ta little Ta signed
.It riscv64sf Ta little Ta signed
.El
.Ss Page Size
.Bl -column -offset indent "Architecture" "Page Sizes"
.It Sy Architecture Ta Sy Page Sizes
.It aarch64 Ta 4K, 2M, 1G
.It amd64 Ta 4K, 2M, 1G
.It armv6 Ta 4K, 1M
.It armv7 Ta 4K, 1M
.It i386 Ta 4K, 2M (PAE), 4M
.It mips Ta 4K
.It mipsel Ta 4K
.It mipselhf Ta 4K
.It mipshf Ta 4K
.It mipsn32 Ta 4K
.It mips64 Ta 4K
.It mips64el Ta 4K
.It mips64elhf Ta 4K
.It mips64hf Ta 4K
.It powerpc Ta 4K
.It powerpcspe Ta 4K
.It powerpc64 Ta 4K
.It riscv64 Ta 4K, 2M, 1G
.It riscv64sf Ta 4K, 2M, 1G
.El
.Ss Floating Point
.Bl -column -offset indent "Architecture" "float, double" "long double"
.It Sy Architecture Ta Sy float, double Ta Sy long double
.It aarch64 Ta hard Ta soft, quad precision
.It amd64 Ta hard Ta hard, 80 bit
.It armv6 Ta hard Ta hard, double precision
.It armv7 Ta hard Ta hard, double precision
.It i386 Ta hard Ta hard, 80 bit
.It mips Ta soft Ta identical to double
.It mipsel Ta soft Ta identical to double
.It mipselhf Ta hard Ta identical to double
.It mipshf Ta hard Ta identical to double
.It mipsn32 Ta soft Ta identical to double
.It mips64 Ta soft Ta identical to double
.It mips64el Ta soft Ta identical to double
.It mips64elhf Ta hard Ta identical to double
.It mips64hf Ta hard Ta identical to double
.It powerpc Ta hard Ta hard, double precision
.It powerpcspe Ta hard Ta hard, double precision
.It powerpc64 Ta hard Ta hard, double precision
.It riscv64 Ta hard Ta hard, quad precision
.It riscv64sf Ta soft Ta soft, quad precision
.El
.Ss Default Tool Chain
.Fx
uses
.Xr clang 1
as the default compiler on all supported CPU architectures,
LLVM's
.Xr ld.lld 1
as the default linker, and
ELF Tool Chain binary utilities such as
.Xr objcopy 1
and
.Xr readelf 1 .
.Ss MACHINE_ARCH vs MACHINE_CPUARCH vs MACHINE
.Dv MACHINE_CPUARCH
should be preferred in Makefiles when the generic
architecture is being tested.
.Dv MACHINE_ARCH
should be preferred when there is something specific to a particular type of
architecture where there is a choice of many, or could be a choice of many.
Use
.Dv MACHINE
when referring to the kernel, interfaces dependent on a specific type of kernel
or similar things like boot sequences.
.Bl -column -offset indent "Dv MACHINE" "Dv MACHINE_CPUARCH" "Dv MACHINE_ARCH"
.It Dv MACHINE Ta Dv MACHINE_CPUARCH Ta Dv MACHINE_ARCH
.It arm64 Ta aarch64 Ta aarch64
.It amd64 Ta amd64 Ta amd64
.It arm Ta arm Ta armv6, armv7
.It i386 Ta i386 Ta i386
.It mips Ta mips Ta mips, mipsel, mips64, mips64el, mipshf, mipselhf, mips64elhf, mipsn32
.It powerpc Ta powerpc Ta powerpc, powerpcspe, powerpc64
.It riscv Ta riscv Ta riscv64, riscv64sf
.El
.Ss Predefined Macros
The compiler provides a number of predefined macros.
Some of these provide architecture-specific details and are explained below.
Other macros, including those required by the language standard, are not
included here.
.Pp
The full set of predefined macros can be obtained with this command:
.Bd -literal -offset indent
cc -x c -dM -E /dev/null
.Ed
.Pp
Common type size and endianness macros:
.Bl -column -offset indent "BYTE_ORDER" "Meaning"
.It Sy Macro Ta Sy Meaning
.It Dv __LP64__ Ta 64-bit (8-byte) long and pointer, 32-bit (4-byte) int
.It Dv __ILP32__ Ta 32-bit (4-byte) int, long and pointer
.It Dv BYTE_ORDER Ta Either Dv BIG_ENDIAN or Dv LITTLE_ENDIAN .
.Dv PDP11_ENDIAN
is not used on
.Fx .
.El
.Pp
Architecture-specific macros:
.Bl -column -offset indent "Architecture" "Predefined macros"
.It Sy Architecture Ta Sy Predefined macros
.It aarch64 Ta Dv __aarch64__
.It amd64 Ta Dv __amd64__, Dv __x86_64__
.It armv6 Ta Dv __arm__, Dv __ARM_ARCH >= 6
.It armv7 Ta Dv __arm__, Dv __ARM_ARCH >= 7
.It i386 Ta Dv __i386__
.It mips Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32
.It mipsel Ta Dv __mips__, Dv __mips_o32
.It mipselhf Ta Dv __mips__, Dv __mips_o32
.It mipshf Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32
.It mipsn32 Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n32
.It mips64 Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64
.It mips64el Ta Dv __mips__, Dv __mips_n64
.It mips64elhf Ta Dv __mips__, Dv __mips_n64
.It mips64hf Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64
.It powerpc Ta Dv __powerpc__
.It powerpcspe Ta Dv __powerpc__, Dv __SPE__
.It powerpc64 Ta Dv __powerpc__, Dv __powerpc64__
.It riscv64 Ta Dv __riscv, Dv __riscv_xlen == 64
.It riscv64sf Ta Dv __riscv, Dv __riscv_xlen == 64, Dv __riscv_float_abi_soft
.El
.Pp
Compilers may define additional variants of architecture-specific macros.
The macros above are preferred for use in
.Fx .
.Ss Important Xr make 1 variables
Most of the externally settable variables are defined in the
.Xr build 7
man page.
These variables are not otherwise documented and are used extensively
in the build system.
.Bl -tag -width "MACHINE_CPUARCH"
.It Dv MACHINE
Represents the hardware platform.
This is the same as the native platform's
.Xr uname 1
.Fl m
output.
It defines both the userland / kernel interface, as well as the
bootloader / kernel interface.
It should only be used in these contexts.
Each CPU architecture may have multiple hardware platforms it supports
where
.Dv MACHINE
differs among them.
It is used to collect together all the files from
.Xr config 8
to build the kernel.
It is often the same as
.Dv MACHINE_ARCH
just as one CPU architecture can be implemented by many different
hardware platforms, one hardware platform may support multiple CPU
architecture family members, though with different binaries.
For example,
.Dv MACHINE
of i386 supported the IBM-AT hardware platform while the
.Dv MACHINE
of pc98 supported the Japanese company NEC's PC-9801 and PC-9821
hardware platforms.
Both of these hardware platforms supported only the
.Dv MACHINE_ARCH
of i386 where they shared a common ABI, except for certain kernel /
userland interfaces relating to underlying hardware platform
differences in bus architecture, device enumeration and boot interface.
Generally,
.Dv MACHINE
should only be used in src/sys and src/stand or in system imagers or
installers.
.It Dv MACHINE_ARCH
Represents the CPU processor architecture.
This is the same as the native platforms
.Xr uname 1
.Fl p
output.
It defines the CPU instruction family supported.
It may also encode a variation in the byte ordering of multi-byte
integers (endian).
It may also encode a variation in the size of the integer or pointer.
It may also encode a ISA revision.
It may also encode hard versus soft floating point ABI and usage.
It may also encode a variant ABI when the other factors do not
uniquely define the ABI (e.g., MIPS' n32 ABI).
It, along with
.Dv MACHINE ,
defines the ABI used by the system.
For example, the MIPS CPU processor family supports 9 different
combinations encoding pointer size, endian and hard versus soft float (for
8 combinations) as well as N32 (which only ever had one variation of
all these).
Generally, the plain CPU name specifies the most common (or at least
first) variant of the CPU.
This is why mips and mips64 imply 'big endian' while 'armv6' and 'armv7'
imply little endian.
If we ever were to support the so-called x32 ABI (using 32-bit
pointers on the amd64 architecture), it would most likely be encoded
as amd64-x32.
It is unfortunate that amd64 specifies the 64-bit evolution of the x86
platform (it matches the 'first rule') as everybody else uses x86_64.
There is no standard name for the processor: each OS selects its own
conventions.
.It Dv MACHINE_CPUARCH
Represents the source location for a given
.Dv MACHINE_ARCH .
It is generally the common prefix for all the MACHINE_ARCH that
share the same implementation, though 'riscv' breaks this rule.
For example,
.Dv MACHINE_CPUARCH
is defined to be mips for all the flavors of mips that we support
since we support them all with a shared set of sources.
While amd64 and i386 are closely related, MACHINE_CPUARCH is not x86
for them.
The
.Fx
source base supports amd64 and i386 with two
distinct source bases living in subdirectories named amd64 and i386
(though behind the scenes there's some sharing that fits into this
framework).
.It Dv CPUTYPE
Sets the flavor of
.Dv MACHINE_ARCH
to build.
It is used to optimize the build for a specific CPU / core that the
binaries run on.
Generally, this does not change the ABI, though it can be a fine line
between optimization for specific cases.
.It Dv TARGET
Used to set
.Dv MACHINE
in the top level Makefile for cross building.
Unused outside of that scope.
It is not passed down to the rest of the build.
Makefiles outside of the top level should not use it at all (though
some have their own private copy for hysterical raisons).
.It Dv TARGET_ARCH
Used to set
.Dv MACHINE_ARCH
by the top level Makefile for cross building.
Like
.Dv TARGET ,
it is unused outside of that scope.
.El
.Sh SEE ALSO
.Xr src.conf 5 ,
.Xr build 7
.Sh HISTORY
An
.Nm
manual page appeared in
.Fx 11.1 .